Communicative Strategy in a Formal Model of Dispute

Mare Koit

Institute of Computer Science, University of Tartu, J. Liivi 2, Tartu, Estonia

Keywords: Dispute, Model, Argument, Communicative Strategy, Communicative Space.

Abstract: We study human-human dialogues in a natural language where the communicative goal of the initiator of

dialogue is to bring the partner to a decision to do a certain action. If the partner does not accept the goal

then dispute will start. Arguments for and against of doing the action will be presented by the participants

and finally, one of them wins and another loses the dispute. We present a formal model of dispute which

includes a model of argument. We discuss involvement of the notion of communicative strategy in the

model. A communicative strategy is considered as an algorithm used by a participant for achieving his or

her communicative goal. A communicative strategy determines also how a participant is moving in

‘communicative space’ during interaction. Communicative space is characterized by a number of

coordinates (e.g. social distance between participants, intensity of communication, etc.). A limited version

of the model of dispute is implemented on the computer.

1 INTRODUCTION

There are two kinds of dialogues which deal with

argumentation: disputes and negotiations (Walton

and Krabbe, 1995). Each dialogue consists of a

sequence of dialogue acts that follow certain patterns

of interaction. When initiating dispute, ”a speaker

asserts a proposition expecting to be asked for

reasons/arguments in support of it and being

prepared to present and defend them” (Wagner,

1998). The initiator of dispute (the proponent) wins

the dispute if the opponent has to accept the initial

assertion. On the contrary, the opponent wins if the

proponent has to withdraw.

When making a proposal or assertion in

negotiation, the proponent, differently as compared

with dispute, has to be prepared to receive critiques

or counterproposals and react to them. “Negotiation

is a form of interaction in which a group of agents

with conflicting interests and a desire to cooperate

try to come to a mutually acceptable agreement on

the division of scarce resources” (Rahwan et al.,

2004). Each party tries to gain an advantage for

themselves. Negotiation is intended to aim at

compromise.

Many researchers have been modelling

negotiation on the computer and investigating

formalization of argument. Good overviews of the

area can be found e.g. in (Chesñevar et al., 2000;

Prakken and Vreeswijk, 2002; Besnard and Hunter,

2008).

We study human-human dialogues in a natural

language between two participants where the

initiator (A) makes a proposal to the partner (B) to

do an action (D) and argues for positive outcomes of

doing D. If B refuses to do D then the participants

have been involved into dispute. Both parties can

present their arguments and counterarguments and

finally, whether A wins, i.e. achieves B’s decision

“do D”, or A loses, i.e. has to withdraw.

We have worked out and implemented on the

computer a formal model of dispute (Koit and Õim,

2014a). In the current paper, we will further develop

the model and concentrate on the ways how

participants can achieve their communicative goals –

communicative strategies. We also introduce the

notion of communicative space. When

communicating, the participants are ‘moving’ in a

communicative space from one ‘point’ to another

and depending on their locations, they choose their

communicative strategies.

The paper has the following structure. In section

2 we introduce our model of dispute including the

notions of communicative strategy and

communicative space. In section 3 we apply the

model to human-human disputes in a natural

language. We investigate a dispute between a sales

clerk of an educational company and a customer

who is supposed to take a training course offered by

the clerk. We study the strategies that the

489

Koit M..

Communicative Strategy in a Formal Model of Dispute.

DOI: 10.5220/0005274904890496

In Proceedings of the International Conference on Agents and Artiﬁcial Intelligence (ICAART-2015), pages 489-496

ISBN: 978-989-758-074-1

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

participants implement for achieving their opposite

communicative goals and their movement in

communicative space. Section 4 discusses the

analysed dispute and proposes a communicative

strategy that can be effective in a computer system

for argumentation. In section 5 we draw conclusions.

2 A MODEL OF DISPUTE

2.1 Components of the Model

Let us consider a dialogue between two participants

A and B (humans or artificial agents) in a natural

language (Koit and Õim, 2013; 2014b). Let the

communicative goal of A be “B decides to do the

action D”. A has a partner model at his disposal – an

image about B – which gives him an opportunity to

believe that B will agree to do the action. In

constructing his first turn, A plans the dialogue acts

(e.g. proposal, request, etc. depending on his partner

model) and determines their verbal form (i.e. the

utterances). The partner B interprets A’s turn and

generating her response, she triggers a reasoning

procedure in her mind in order to make a decision –

to do D or not. In the reasoning process, B weighs

her resources, positive and negative outcomes of

doing D and finally, she makes a decision. Then she

in her turn plans the dialogue acts (e.g. agreement,

refusal, refusal with argument, etc.) and their verbal

form in order to inform A about her decision. If B

agrees to do D then the dialogue finishes (A has

reached his communicative goal). If B’s response is

refusal then A must change his partner model (which

did not correspond to the reality because A supposed

that B will agree to do D) and he has to find out new

arguments to convince B to make the decision.

B can add arguments to her refusal. These

(counter)arguments give information to A about the

reasoning process that brought B to the (negative)

decision.

Then A has to find a new argument for doing D

by B and the process continues in a similar way.

2.1.1 Reasoning Model

Our reasoning model consists of two parts: (1) a

model of human motivational sphere; (2) reasoning

procedures (Koit and Õim, 2014b). In the

motivational sphere of a reasoning person some

basic factors can be found that regulate reasoning

about doing an action D. We call these factors

WISH-, NEEDED- and MUST-determinants,

respectively. First, a subject has a wish to do D if the

pleasant aspects of D outweigh the unpleasant ones;

second, doing D is needed for the subject if the

useful aspects of D outweigh the harmful ones; and

third, a subject must to do D if not doing implies

some punishment.

We represent the model of motivational sphere

of a subject by the following vector of ‘weights’

(with numerical values of its components):

w = (w(resources), w(pleasant),

w(unpleasant), w(useful), w(harmful),

w(obligatory), w(prohibited),

w(punishment-do), w(punishment-not)).

In the description, w(pleasant), etc. mean the

weight of pleasant, etc. aspects of D; w(punishment-

do) – weight of punishment for doing D if it is

prohibited, and w(punishment-not) – weight of

punishment for not doing D if it is obligatory. Here

w(resources) = 1, if subject has the resources

necessary to do D (otherwise 0, i.e. we suppose that

all the needed resources either exist or not);

w(obligatory) = 1, if D is obligatory for the

reasoning subject (otherwise 0); w(prohibited) = 1, if

D is prohibited (otherwise 0). The values of other

weights can be non-negative natural numbers.

Although in reality people do not operate with

numbers but use words for characterising different

aspects of an action (e.g. extremely useful, not much

useful, neither useful nor harmful, etc.), the

existence of certain scales also in human everyday

reasoning is apparent.

The second part of the reasoning model consists

of reasoning procedures that supposedly regulate

human action-oriented reasoning. A reasoning

procedure depends on the determinant which

triggers it (in our model, WISH, NEEDED or

MUST). Every reasoning procedure represents steps

that the subject goes through in his/her reasoning

process; these consist in computing and comparing

the summarized weights of different aspects of D;

and the result is the decision: to do D or not (Koit

and Õim, 2013).

We use two different vectors of weights in our

model of dispute: w

(B’s idiosyncratic model which

represents B’s actual evaluations of D’s aspects) and

(the partner model – A’s beliefs concerning B’s

evaluations).

2.1.2 Communicative Space

Communication between two participants can be

different: collaborative or confrontational, polite or

impolite, friendly or unfriendly, etc. Moreover, a

dialogue which has started violently can finish

peacefully if the participants achieve a compromise,

ICAART2015-InternationalConferenceonAgentsandArtificialIntelligence

490

therefore, the character of communication can

change.

Healey et al. (2008) declare that “there are

important differences in the quality of human

interaction – in degrees of interpersonal, as opposed

to physical, closeness – that are important for the

organization of human activities and, consequently,

for design”. They suppose that the concept of

communication space provides a useful approach to

thinking about the basic organization of human

interaction (see also Brown and Levinson, 1999).

We can imagine ‘communicative space’ where

the participants are ‘moving’ from one ‘point’ to

another. Communicative space can be represented as

an n-dimensional space (n > 0) where the different

coordinates characterize the different features of

communication. We are able to specify at least the

following features: social distance between

participants (on the scale from near to far, or, in

other words, from familiar to remote), cooperation

(from collaborative to confrontational), politeness

(from polite to impolite), personality (from personal

to impersonal), modality (from friendly to hostile),

intensity (from peaceful to vehement). We suppose

that just as in the case of human motivational sphere,

people have an intuitive ‘theory’ of these

coordinates.

The values of the coordinates can be expressed

by specific words like in case of pleasant, useful,

etc. aspects of an action D. Instead, we use

numerical values as approximations in our model

and represent the values of coordinates of the

communicative space as -1, 0 or +1. For example,

the value -1 on the scale of social distance means

that the participant is “far” from his/her partner. The

value 0 marks “neutral” and the value +1 “near”

social distance. In the same time, the partner when

communicating can express a different social

distance (e.g. “neutral” instead of “far”), i.e. it is

asymmetrical.

A feature vector can be assigned to each

utterance of a dialogue that determines the point in

the communicative space where the author of the

utterance is just located. In the following example

(start of a phone call), the friends Siiri and Marju are

located in the same communication point which can

be represented by the vector (+1,+1,+1,+1,+1,+1),

i.e. the social distance is “near”, interaction is

“cooperative”, “polite”, “personal”, “friendly”, and

“peaceful”:

Siiri: Marju?

Marju: Ciao.

2.1.3 Communicative Strategies and Tactics

A communicative strategy is an algorithm used by a

participant for achieving his/her goal in the

interaction (Fig.1; Koit and Õim, 2014b).

1) Choose an initial communication

point in the communicative space.

2) Choose communicative tactics.

3) Implement the tactics to generate an

utterance: inform the partner of the

communicative goal (agreeing to do an

action D).

4) Did the partner agree to do D? If

yes then finish (the communicative goal

has been achieved).

5) Did the partner postpone the

decision? If yes then finish (the

communicative goal has not been

achieved but can be achieved in the

future).

6) Give up? If yes then finish (the

communicative goal has not been

achieved).

7) Change the point in the

communicative space? If yes then choose

a new point.

8) Change the communicative tactics? If

yes then choose new tactics.

9) Implement the tactics to generate an

utterance (argument).

10)Go to 4.

Figure 1: Communicative strategy of the initiator of

dialogue.

The initiator A can realize his communicative

strategy in different ways: entice, persuade or

threaten the partner B to do D (respectively, stress

pleasantness or usefulness of D or punishment for

not doing D if it is obligatory). We call these ways

of realization of a communicative strategy

communicative tactics.

The partner B uses a similar communicative

strategy. The only difference is that B does not have

initiative at the beginning of the dialogue and her

communicative goal is opposite as compared with

A’s one: “do not do D”. Similarly, B can use various

communicative tactics realizing her communicative

strategy, e.g. defence or attack. In the last case, she

takes the initiative and proposes arguments against

doing D trying to turn off A’s arguments for doing

the action.

2.2 The Structure of Dispute

As we suppose, A and B have contradictory goals

when starting interaction (dispute). A’s

communicative goal is “B will do D”, B’s goal is “B

will not do D”. We suppose in our model that both A

and B can use a common set of reasoning

CommunicativeStrategyinaFormalModelofDispute

491

algorithms. We also suppose that both A and B have

fixed sets of dialogue acts and corresponding

utterances which are classified semantically, for

example, P

increasing_pleasantness

for stressing pleasantness

of D, P

decreasing_harmfulness

for downgrading

harmfulness, etc. for A and P

decreasing_pleasantness

increasing_harmfulness

, etc. for B.

Starting a dispute, A fixes a partner model w

and determines the communicative tactics T which

he will use, i.e. he accordingly fixes a reasoning

algorithm R which he will try to trigger in B’s mind.

B has her own model of motivational sphere – w

She determines a reasoning procedure R

which she

will use in order to make a decision about doing D.

The general structure of A’s argument is as

follows (cf. Besnard and Hunter, 2008; Koit and

Õim, 2014a):

<{R, T, w

, proposition

}, claim

where

 R

is the reasoning algorithm which A is trying to

trigger in B,

 T is the communicative tactics used,

 w

= (w

(resources), w

(pleasant), w

(unpleasant), w

(useful), w

(harmful),

(obligatory), w

(prohibited),

(punishment-do), w

(punishment-not)) is

the current partner model (at time i),

 proposition

denotes the utterance chosen by A in

order to influence one of the weights in the

partner model, after what R will supposedly give

B’s positive decision (do D) on the changed

model w

i+1

(at time i+1); its weight is

w(proposition

 claim

= “B decides to do D“ (A’s

communicative goal).

The structure of B’s (counter)argument is

analogous, with the difference that w

is used

instead of w

and claim

is “B does not do D” (i.e.

B’s communicative goal).

In dispute, only propositions incorporated in

arguments are explicitely presented by participants,

the other components of arguments are implicit.

The general structure of dispute is given in Fig. 2

(the dialogue acts in parentheses can be missed).

Both A and B can indicate that the finishing

conditions are fulfilled: (1) the communicative goal

is already achieved, (2) give up regardless of having

new arguments, (3) there are no arguments to

continue the fixed tactics but no new tactics will be

chosen regardless of having some tactics not

implemented so far, (4) all the tactics are already

implemented and all the possible arguments are used

without achieving the communicative goal.

A: proposal (+ argument(s))

REPEAT

(

B: question

A: answer/giving information

)

B: refusal (+ argument(s))

(

A: question

B: answer/giving information

)

A: argument(s)

UNTIL finishing conditions are fulfilled

Figure 2: Structure of dispute.

If B gives up then she makes the decision to do D

and A has achieved his communicative goal (A wins

and B loses the dispute). If A gives up then he does

not achieve his communicative goal and B will not

do D (A loses and B wins). If B postpones her

decision at the end of dialogue then there are neither

winners nor losers.

Questions are asked by participants in order to

make choices among different propositions that can

be used in argumentation.

We have implemented the model of dispute as a

computer program, which, as we believe, can be

used for training argumentation skills (Koit and

Õim, 2014b). The user who plays the role of B

interacts with the computer (A) in written Estonian,

either choosing ready-made sentences as

counterarguments against performing the action

proposed by A or putting in free texts. In the last

case, cue words are used by the computer in order to

understand user sentences. The computer has ready-

made sentences for stressing or downgrading the

values of different aspects of the proposed action

depending on the user model. So far, we have not

used communicative space in the implementation

and did not consider movement of participants from

one communication point to another during

interaction.

3 HUMAN-HUMAN DISPUTE

In this section, we apply the introduced model to

actual human-human disputes. Our special interest is

to study communicative strategies used by

participants for achieving their communicative goals

and the communicative space where they are moving

during interaction.

We have analysed a sub-corpus of 52

telemarketing calls of the Estonian dialogue corpus

(Hennoste et al., 2008). The calls are recorded in

authentic situations and transcribed using the

ICAART2015-InternationalConferenceonAgentsandArtificialIntelligence

492

transcription of Conversation Analysis (Hutchby and

Wooffitt, 1998). In the dialogues, sales clerks of an

educational company (which changed name is

Tiritamm) are calling to potential customers (who

are managers or personnel officers of other

institutions). Sales clerks offer training courses

(language, marketing, business training, etc.) for

employees of these institutions.

The communicative goal of a sales clerk (A) is

that the customer (B) will decide to buy a training

course (the action D). The communicative goal of B

is usually opposite: do not take the course because it

spends time and money (although the courses will

still be useful for B). Therefore, the participants are

involved into dispute.

Typically, a lot of calls are needed before a

customer makes her final decision. In our analysed

sub-corpus, the calls belong to the beginning period

of negotiations. This is the reason why the customers

mostly postpone their decisions and there are no

winners or losers in the disputes. The final decision

(to take a course or not) has been made only in seven

dialogues.

A telemarketing call like all institutional calls

has two conventional parts at the beginning and at

the end, respectively – greeting and leave-taking.

Negotiation takes place and a decision will be made

in the main part of a call. Several phases can be

differentiated in the main part (Koit, 2014): (1)

opening – A introduces himself and his educational

company and makes sure that B is the requested

person who is responsible for education of the

employees of her institution, (2) finding out B’s

needs where A asks several questions about B’s

institution and collects information from B’s

answers, (3) argumentation for and against of a

training course, (4) B’s decision. The phases (2) and

(3) can alternate or (2) can miss if the participants

have previously been in contact and A already has

sufficient knowledge about B’s institution.

Opening a dialogue, A determines an initial point

in communicative space: as an official person, he

typically chooses a neutral social distance (value 0

on the scale [-1, +1], collaborative and polite

interaction (values +1), neutral personality, modality

and intensity (all values 0).

In the case if the same participants have already

had previous conversations, A can choose a shorter

social distance (+1) and greater personality (+1). In

the following example, A moves away from the

neutral, official position and is interested in B’s

person:

A: mt (0.2) kuidas on elu ´vahepeal

läinud, kõik kenad ´reisid on

´seljataha [jäänud.]

How have you been doing in the meanwhile, have you

past all the nice trips?

During a dialogue, A typically stands in the

initially fixed communication point while B can

change the value of any coordinate of

communicative space. A implements the

communicative tactics of persuasion indicating and

stressing usefulness of the offered course for B. The

other possible tactics (enticement, threatening) are

excluded because A is an official person and has to

keep himself in check.

Let us consider the following example where a

sales clerk A is calling to the chairperson B of a

furniture salon. A starts collaborative and polite

interaction (values +1) and fixes values 0 for other

coordinates of the communicative space. The partner

B in her turn determines her initial communication

point. Similarly with A, she also starts collaborative

and polite interaction but differently, she chooses the

values +1 for cooperation and modality as indicated

by the comment ((kindly and friendly)) in the

following transcript:

A: õhtust.

Good aftenoon.

palun Liisi ´Harvsõlg.

I would like to speak with Liisi Harvsõlg.

B: jaa=ma ´kuulen.

Yes, I’m listening.

((kindly and friendly))

A continues in the same communication point:

A: .hhh ee minu nimi on ´Maanus ´Kriisa

ja elistan ´Tiritamm < ´Eestist. >

My name is Maanus Kriisa and I’m calling from Tiritamm

Estonia.

.hh kas > ´Tiritamme=nimi < on teile

´tuttav.

Are you acquainted with Tiritamm?

By replying, B suddenly jumps to another

communication point changing the values of social

distance, cooperation and modality (new values -1)

as indicated by the comment ((unfriendly)):

B: ei ausalt=öelda ´küll ei=ole.

No, truly.

((unfriendly))

After that A, keeping the chosen communication

point and communicative tactics, introduces his

company, describes its activities and then

(indirectly) proposes a training course which deals

with relations between servicemen and clients and

should be useful for B:

CommunicativeStrategyinaFormalModelofDispute

493

/---/

A: et=ee (.) .hh nimelt Tiritamme

tegeleb sis ´kliendi´suhete ee ütleme

´inimeste: ´arendamisega

´kliendi´suhete > ´valdkonnas=se

tähendab sis < < ´teeninduse ´müügi >

.h ´turunduse=ja=´juhtimise

´valdkonnas.

Tiritamm is considering relations with clients and

education of people in service and sale.

(.) .hh ja: minu ´küsimus oligi see=et

kas te peate: vajalikuks just ´nendes

´valdkondades ´inimeste ´arendamist=

Is it important for you to educate your people in these

fields?

/---/

B does not shorten the social distance nor

decrease antagonism. She presents a lot of

counterarguments:

B: [{-} kuulge] eeq ´te=olete=küll ´nii

vähe ´valesse kohta sattund.

meil on ´kolm inimest ´tööl.

Look here! You are ever calling into a wrong place. We

have only three employees.

((unfriendly))

/---/

.hhh põhjus on lihtsalt selles et minu

mõlemad müüja=konsultandid nad ´väga

hästi haritud.

The reason is simply that both my sellers-consultants are

educated very well.

/---/

$ mul on ´topelt ´kõrgharidus. $

I myself have double higher education.

A explains why the course is useful for B:

A: /---/ hh aga kui on teil ütleme

eksklu´ssiivne ´toode ta=on (.) ilmselt

´inglise ´mööbel ta on=nagu te

´ütlesite ´ka (.) väg- väga

´kalli´hinnalne.

So you have an exclusive product, obviously English

furniture, it is very expensive as you said.

.hh et=ee sis on seal ilmselt ´oluline

ka see ´väga:q ee ´hea ´teeninduse

´tase mis teil [ka ilmselt] ´on.

Then a very high level of service is necessary, what you

obviously have.

/---/

.hhh et ´tehakse sisuliselt

presentatsi´oo[ne kus siis=ee

Thus, you do presentations.

B: [´just ´nimelt.]

Exactly.

/---/

A: .hhhh ?jaa?jaa.

.h SEDA PAKUB KA ´TIRITAMM muidu

´välja: ee kuidas siis=õõ ´läbi [viia

´presentatsi´oone.]

Tiritamm also offers a course how to do presentations.

B gives more counterarguments and refuses to

take the course:

A: /---/ nii=et=te ei pea ´vajalikuks

inimeste ´arendamist nendes

´vald[kondades.]

Then you believe that education of people is not necessary

in these fields.

B: [tändap=ee] ma ´kordan me saame ´ise

selle teemaga ´hakkama.

I tell once more, we can make it ourselves.

In this dispute, A fails to achieve his

communicative goal (he loses). He does not change

the communication point and communicative tactics

chosen by him at the beginning of conversation. B

has changed her communication point directly after

the first turn. She also keeps her new communication

point and the communicative tactics during

conversation. The participants do not reach an

agreement. Still, at the end of conversation, A offers

to B to send a catalogue with a list of training

courses offered by his educational company. B is not

against. In this way, A can hope that it will be

possible for him to win a dispute with B in the

future.

4 DISCUSSION

The corpus analysis demonstrates that the model

introduced in section 2 can be used for description of

actual human-human disputes. We can consider

communicative strategies implemented by

participants for achieving their communicative

goals. We can also see how the features of

communication are changing during interaction, i.e.

how the participants are behaving in communicative

space.

A question arises how to recognise

communication points automatically? Assuming that

transcripts of spoken dialogues are available, we can

analyse comments (in our used transcription system,

in double parentheses) as demonstrated by the

examples of the previous section, e.g. ((unfriendly)).

Another way is to extract information from the

header of a transcript. Each transcript in our corpus

is provided with a header that lists situational

factors, among them relations between participants

(strangers/ acquaintances/ intimates), status (equal/

lower/ higher), gender, age, etc. (Hennoste et al.,

2008). But when modelling debate on the computer

ICAART2015-InternationalConferenceonAgentsandArtificialIntelligence

494

and aiming to recognise a communication point

where the user stands, the computer could extract

opinions and emotions expressed by user utterances.

In the case of spoken interaction, the features of

speech can be used in addition. Still, automatic

recognition of coordinates of communicative space

remains for the further work.

The most important phase of a telemarketing call

is a clerk’s argumentation for taking a training

course. Arguments of sales clerks are presented as

assertions and customers can accept or reject them.

In our analysed dialogues, the customers typically

accept the assertions of clerks – it shows that the

clerks succeed to choose the ‘right’ arguments (Koit,

2014). Still, B’s accept is usually followed by a

counterargument. The argumentation chain looks

A: argument

– B: accept

+ counterargument

…

A: argument

– B: accept

+ counterargument

The situation is different when B is driving to a

negative decision (as in the example analysed in the

previous section). In this case, B does not accept A’s

arguments and takes the initiative starting to present

assertions/counterarguments herself. A always

accepts B’s assertions but he still provides his

arguments as additional information. The

argumentation chain looks like

A: argument

– B: reject

+ counterargument

A: accept

+ argument

– B: reject

counterargument

…

A: accept

+ argument

– B: reject

counterargument

Therefore, we can see that the general structure

of actual dialogues corresponds to the model of

dispute presented in section 2.

What can we learn from the analysis of human-

human disputes for implementation of dispute on the

computer? A good way seems to follow the sales

clerks’ strategies: take and hold the initiative and

propose ’obvious’ arguments for the requested

action – the statements that do not provoke the

partner’s rejection but accept. In order to have such

arguments at disposal, it is necessary to know as

possible more about the partner in relation to the

requested action. That is the reason why explanation

of the customer’s needs is a necessary phase in our

analysed telemarketing calls. Still, both a sales clerk

and a customer are restricted when communicating

because both they are official persons who represent

their institutions and therefore, have to play certain

roles. A sales clerk who is interested in selling

training courses has to keep a fixed communication

point and a fixed strategy. A customer has more

freedom: she may defend her negative decision or

also attack the clerk’s proposal. She may vary her

features of communication, e.g. change the intensity

from peaceful to vehement or go over to

confrontation instead of collaboration.

Our dialogue corpus does not yet contain more

different kinds of disputes although it would be

interesting to look for communicative strategies

expressed e.g. in quarrels and to study how the

participants are moving in the communicative space

when having a quarrel. Still, the scenario where the

computer plays the role of an official person who

behaves cooperatively, politely, friendly, etc. is

more realistic for implementation.

5 CONCLUSIONS

We study dialogues in a natural language where one

participant (initiator of interaction, A) has a

communicative goal that the partner (B) will decide

to do an action D. If B’s communicative goal is

opposite (“do not do D”) then the participants are

involved into dispute. When reasoning about doing

D, B considers different positive and negative

aspects of D. If the positive aspects weigh more than

negative then the decision will be “do D”. On the

contrary, if the negative aspects weigh more the

decision will be “do not do D”. Initiator A chooses a

suitable communicative strategy in order to

influence B’s reasoning and achieve the positive

decision: he stresses positive and downgrades

negative aspects of doing D. Different arguments for

doing D are presented in a systematic way, e.g. A

stresses time and again usefulness of D. Still, if B

takes over the initiative then A can also act

passively, only averting the (counter)arguments

presented by B and not stressing any positive aspect

of the action.

When interacting, the participants are moving in

communicative space which can be characterized by

a number of features (coordinates) such as social

distance between the partners (far between

adversaries, near between friends), intensity of

communication (peaceful, vehement), etc. We

represent values of the coordinates as -1, 0 or +1.

We analyse human-human telemarketing calls

where a sales clerk of an educational company

proposes training courses to a customer who

typically does not want to buy any course. When

starting dispute, the sales clerk determines an initial

CommunicativeStrategyinaFormalModelofDispute

495

point in communicative space and a certain way to

realize the communicative strategy (tactics) and

retains them during a dialogue. The customer can

change her strategy and also move from one

communication point to another during

conversation.

We have implemented the model of dispute as a

computer program where the computer plays A’s and

the user B’s role. So far, the implementation does

not include the formalisation of communicative

space. This remains for the further work.

ACKNOWLEDGEMENTS

This work was supported by the European Regional

Development Fund through the Estonian Centre of

Excellence in Computer Science (EXCS), and the

Estonian Research Council (projects IUT-2056,

ETF9124, and ETF8558).

REFERENCES

P. Besnard, and A. Hunter. 2008. Elements of

Argumentation. Cambridge, MA: MIT Press.

P. Brown, and S. C. Levinson. 1999. Politeness: Some

universals in language usage. In The Discourse

Reader. A. Jaworski and N. Coupland, Eds. London:

Routledge, 321–335.

C. Chesñevar, A. Maguitman, and R. Loui. 2000. Logical

models of argument. In ACM Computing Surveys,

32(4), 337–383.

P. G. T. Healey, G. White, A. Eshghi, A. J. Reeves, and A.

Light. 2008. Communication spaces. In Computer

Supported Cooperative Work. Springer, 17:169–193.

DOI: 10.1007/s10606-007-9061-4.

T. Hennoste, O. Gerassimenko, R. Kasterpalu, M. Koit, A.

Rääbis, and K. Strandson. 2008. From human

communication to intelligent user interfaces: corpora

of spoken Estonian. In Proc. of the 6th International

Language Resources and Evaluation (LREC'08).

Marrakech, Morocco: European Language Resources

Association (ELRA), 2025–2032.

I. Hutchby, and R. Wooffitt. 1998. Conversation Analysis.

Principles, Practices and Applications. Cambridge,

UK: Polity Press.

M. Koit. 2014. Towards automatic recognition of the

negotiation strategies: Analysis of human-human

dialogues. In 2014 IEEE International Symposium on

Innovations in Intelligent Systems and Applications

(INISTA 2014) Proceedings. Alberobello, Italy:

Institute of Electrical and Electronics Engineers

(IEEE), 170–176.

M. Koit, and H. Õim. 2014a. Modelling debates on the

computer. In Proceedings of the 6th International

Conference on Knowledge Engineering and Ontology

Development (KEOD 2014). B. Duval, J. van den

Herik, S. Loiseau and J. Filipe, Eds. Rome, Italy:

SciTePress, 361–368.

M. Koit, and H. Õim. 2014b. A computational model of

argumentation in agreement negotiation processes. In

Argument & Computation, 5 (2-3), 209–236, Taylor &

Francis Online. DOI: 10.1080/19462166.2014.915233.

M. Koit, and H. Õim. 2013. From study of human-human

dialogues to reasoning model. Conversational agent in

argumentation dialogue. In: Proc. of the 5th

International Conference on Agents and Artificial

Intelligence (ICAART 2013). J. Filipe and A. Fred,

Eds. Barcelona, Spain: SciTePress, 210–216.

H. Prakken, and G.A.W. Vreeswijk 2002. Logics for

defeasible argumentation. In Handbook of

Philosophical Logic. D.M. Gabbay and F. Guenthner,

Eds. Dordrecht/Boston/London: Kluwer Academic

Publishers. 2nd edition, Vol 4, 219–318.

I. Rahwan, S. D. Ramchurn, N. R. Jennings, P. Mcburney,

S. Parsons, and L. Sonenberg. 2004. Argumentation-

based negotiation. In The Knowledge Engineering

Review. Cambridge University Press. Vol. 18:4, 343–

375. DOI: 10.1017/S0269888904000098.

G. Wagner. 1998. Foundations of Knowledge Systems with

Applications to Databases and Agents. Kluwer

Academic Publishers.

D. Walton, and E. C. W. Krabbe. 1995. Commitment in

Dialogue. Albany: SUNY Press.

ICAART2015-InternationalConferenceonAgentsandArtificialIntelligence

496